The development of human carcinomas has been shown to arise from the accumulation of genetic changes involving both positive regulators of cell function (oncogenes) and negative regulators (tumour suppressor genes). For a normal somatic cell to evolve into a metastatic tumour it requires changes at the cellular level, such as immortalisation, loss of contact inhibition and invasive growth capacity, and changes at the tissue level, such as evasion of host immune responses and growth restraints imposed by surrounding cells, and the formation of a blood supply for the growing tumour.
Molecular genetic studies of colorectal carcinoma have provided substantial evidence that the generation of malignancy requires the sequential accumulation of a number of genetic changes within the same epithelial stem cell of the colon. For a normal colonic epithelial cell to become a benign adenoma, progress to intermediate and late adenomas, and finally become a malignant cell, inactivating mutations in tumour suppressor genes and activating mutations in proto-oncogenes are required (Fearon and Vogelstein, 1990).
The employment of a number of techniques, such as loss of heterozygosity (LOH), comparative genomic hybridisation (CGH) and cytogenetic studies of cancerous tissue, all of which exploit chromosomal abnormalities associated with the affected cell, has aided in the identification of a number of tumour suppressor genes and oncogenes associated with a range of tumour types.
In one aspect, studies of cancers such as retinoblastoma and colon carcinoma have supported the model that LOH is a specific event in the pathogenesis of cancer and has provided a mechanism in which to identify the cancer causing genes. This model is further highlighted in Von Hippel-Lindau (VHL) syndrome, a rare disorder that predisposes individuals to a variety of tumours including clear cell carcinomas of the kidneys and islet cell tumours of the pancreas. Both sporadic and inherited cases of the syndrome show LOH for the short arm of chromosome 3 and somatic translocations involving 3p in sporadic tumours, and genetic linkage to the same region in affected families has also been observed. The VHL tumour suppressor gene has since been identified from this region of chromosome 3 and mutations in it have been detected in 100% of patients who carry a clinical diagnosis of VHL disease. In addition, the VHL gene is inactivated in approximately 50–80% of the more common sporadic form of renal clear cell carcinoma.
The genetic determinants involved in breast cancer are not as well defined as that of colon cancer due in part to the histological stages of breast cancer development being less well characterised. However, as with colon carcinoma, it is believed that a number of genes need to become involved in a stepwise progression during breast tumourigenesis.
Certain women appear to be at an increased risk of developing breast cancer. Genetic linkage analysis has shown that 5 to 10% of all breast cancers are due to at least two autosomal dominant susceptibility genes. Generally, women carrying a mutation in a susceptibility gene develop breast cancer at a younger age compared to the general population, often have bilateral breast tumours, and are at an increased risk of developing cancers in other organs, particularly carcinoma of the ovary.
Genetic linkage analysis on families showing a high incidence of early-onset breast cancer (before the age of 46) was successful in mapping the first susceptibility gene, BRCA1, to chromosome 17q21 (Hall et al., 1990). Subsequent to this, the BRCA2 gene was mapped to chromosome 13q12–q13 (Wooster et al., 1994) with this gene conferring a higher incidence of male breast cancer and a lower incidence of ovarian cancer when compared to BRCA1.
Both BRCA1 and BRCA2 have since been cloned (Miki et al., 1994; Wooster et al., 1995) and numerous mutations have been identified in these genes in susceptible individuals with familial cases of breast cancer.
Additional inherited breast cancer syndromes exist, however they are rare. Inherited mutations in the TP53 gene have been identified in individuals with Li-Fraumeni syndrome, a familial cancer resulting in epithelial neoplasms occurring at multiple sites including the breast. Similarly, germline mutations in the MMAC1/PTEN gene involved in Cowden's disease and the ataxia telangiectasia (AT) gene have been shown to confer an increased risk of developing breast cancer, among other clinical manifestations, but together account for only a small percentage of families with an inherited predisposition to breast cancer.
Somatic mutations in the TP53 gene have been shown to occur in a high percentage of individuals with sporadic breast cancer. However, although LOH has been observed at the BRCA1 and BRCA2 loci at a frequency of 30 to 40% in sporadic cases (Cleton-Jansen et al., 1995; Saito et al., 1993), there is virtually no sign of somatic mutations in the retained allele of these two genes in sporadic cancers (Futreal et al., 1994; Miki et al., 1996). Recent data suggests that DNA methylation of the promoter sequence of these genes may be an important mechanism of down-regulation. The use of both restriction fragment length polymorphisms and small tandem repeat polymorphic markers has identified numerous regions of allelic imbalance in breast cancer suggesting the presence of additional genes, which may be implicated in breast cancer. Data compiled from more than 30 studies reveals the loss of DNA from at least 11 chromosome arms at a frequency of more than 25%, with regions such as 16q and 17p affected in more than 50% of tumours (Devilee and Cornelisse, 1994; Brenner and Aldaz, 1995). However only some of these regions are known to harbour tumour suppressor genes shown to be mutated in individuals with both sporadic (TP53 and RB genes) and familial (TP53, RB, BRCA1, and BRCA2 genes) forms of breast cancer.
Cytogenetic studies have implicated loss of the long arm of chromosome 16 as an early event in breast carcinogenesis since it is found in tumours with few or no other cytogenetic abnormalities. Alterations in chromosome 1 and 16 have also been seen in several cases of ductal carcinoma in situ (DCIS), the preinvasive stage of ductal breast carcinoma. In addition, LOH studies on DCIS samples identified loss of 16q markers in 29 to 89% of the cases tested (Chen et al., 1996; Radford et al., 1995). In addition, examination of tumours from other tissue types have indicated that 16q LOH is also frequently seen in prostate, liver, ovarian and primitive neuroectodermal carcinomas. Together, these findings suggest the presence of a gene mapping to the long arm of chromosome 16 that is critically involved in the early development of a large proportion of breast cancers as well as cancers from other tissue types, but to date no such gene has been identified.